Relational Frame Theory, widely known as RFT, is a behavioral science account of human language and cognition that has increasingly shaped how Applied Behavior Analysis (ABA) is practiced, particularly in autism intervention and education. Where traditional ABA programming often focuses on teaching specific skills one at a time through direct reinforcement, RFT provides a framework for understanding how people learn to relate things to one another in flexible, generative ways and then use those relations to derive new knowledge they were never explicitly taught. The overlap between the two fields is substantial but also contested, with many practicing behavior analysts still finding RFT difficult to translate into everyday clinical work.
What RFT Actually Explains
At its simplest, RFT is a theory about how humans learn language and think abstractly. The central claim is that once you learn certain types of relations between things, you can derive new relations you were never directly taught. For example, if a child learns that a spoken word refers to a particular object, and then learns that another word means the same as the first word, the child can figure out that the second word also refers to that object without anyone having to teach that connection explicitly. RFT calls this “derived relational responding,” and it treats this ability as the foundation of virtually all complex human behavior, from reading comprehension to metaphor to empathy.
This matters for ABA because traditional behavioral approaches tend to teach each response individually. If you want a child to match word A to picture B, you teach that. If you want them to match picture B to word A (the reverse direction), you teach that separately. RFT argues that with the right kind of training, you can establish the conditions for those reverse and extended relations to emerge on their own. The practical payoff is efficiency: instead of drilling every single connection, you train a smaller set of relations and the learner derives the rest.
Both RFT and the more traditional Skinnerian analysis of verbal behavior share roots in behavior analysis, but they developed largely independently. RFT emerged primarily from clinical psychology research, while Skinner’s verbal behavior framework was more closely tied to education and developmental work.
How RFT Changes What Happens in a Therapy Session
In a traditional ABA session for a child with autism, a therapist might teach labels (saying “dog” when shown a dog), requests (asking for a cookie), and imitation (copying an action). These are important building blocks, and Skinner’s classification of verbal behavior into categories like mands, tacts, and intraverbals gives therapists a clear way to identify what a child can and cannot do and to target specific gaps.
RFT adds another layer. Rather than just teaching a child to label or request, an RFT-informed therapist also works on the child’s ability to relate things to each other using contextual cues. That might mean teaching a child that one thing is “the same as” another, that something is “more than” or “less than” something else, or that one event comes “before” or “after” another. The idea is that once these relational repertoires are solid, the child becomes capable of generating novel language and solving problems they have never encountered before. Researchers who advocate integrating both approaches argue that Skinner’s framework is well-suited for identifying early verbal skills and their sources of control, while RFT provides a systematic way to program for generative language through what it calls multiple exemplar training.
Curricula That Put RFT Into Practice
One of the most concrete ways RFT has entered ABA practice is through structured curricula. The PEAK Relational Training System is the most prominent example. PEAK was designed as both an assessment and a treatment protocol for language and cognitive skills in children with autism. It contains four modules: Direct Training and Generalization, which follow a more traditional behavioral framework, and Equivalence and Transformation, which are explicitly built on RFT principles.
Research on PEAK has been encouraging. In one study of young adults with autism, participants mastered target skills during training, and their scores on a separate, well-known developmental assessment (the VB-MAPP) showed an upward trend. Crucially, when training was discontinued, the VB-MAPP scores held steady, and participants even showed mastery of verbal skills that had never been directly targeted, suggesting the kind of generative learning RFT predicts.
The Relational Abilities Index+ (RAI+) is another tool grounded in RFT. Rather than measuring traditional IQ-style abilities, it assesses performance across relational skills like same/opposite, more/less, before/after, and analogy. Validation research found that the RAI+ correlates well with established intelligence measures and numeracy assessments, though not as strongly with educational attainment.
What Outcome Studies Show
The question most parents and practitioners want answered is straightforward: does an RFT-informed approach actually produce better results? The evidence, while still growing, points in a positive direction. A randomized controlled trial comparing autistic children who received ABA therapy with RFT-based content to those who received treatment as usual found that the RFT group gained more skills and showed greater improvements in intelligence scores. Autism severity also appeared to be influenced by which group children were assigned to, suggesting that the type of content within ABA therapy matters, not just the number of hours.
In the educational domain, a study of school-aged children who received an RFT-based intervention targeting relational skills found that the intervention group performed significantly better than controls in both reading and mathematics after the program ended. Gains were large and significant for the treatment group. The control group also improved modestly, but the gap was clear.
These findings are promising, but the field is still in a relatively early stage when it comes to large-scale replication. Most studies involve small samples, and the research base is far more developed on the basic science side of RFT than on its clinical application. A review of all articles on derived stimulus relations published in a major applied journal between 1992 and 2009 concluded that the published work suggested the possibility of applied technologies but largely left the hard work of building those technologies to future researchers.
Perspective-Taking and Social Skills
One of the more intriguing applications of RFT in ABA involves perspective-taking, a skill that many autistic individuals find challenging. RFT frames perspective-taking as a type of relational responding called “deictic” relations. These involve understanding shifts between I/you, here/there, and now/then. If you can flexibly relate your own perspective to someone else’s, in a different place, at a different time, you have the building blocks of empathy and social understanding.
A scoping review of behavior-analytic studies on deictic relational responding and perspective-taking in autistic individuals found seven studies that met inclusion criteria. The review noted some variation in how perspective-taking was tested and trained through deictic relations, which reflects both the novelty of the approach and the lack of a single standardized protocol. Still, the fact that perspective-taking can be framed as a trainable relational skill rather than an all-or-nothing cognitive ability is a meaningful conceptual shift. It suggests that difficulties with social cognition are not fixed limitations but gaps in a learnable repertoire.
Categorization, Hierarchy, and Advanced Language
Beyond basic same/different relations, RFT researchers have tackled more complex cognitive operations like hierarchical classification. In everyday language, this is the ability to understand that a poodle is a type of dog, a dog is a type of animal, and an animal is a type of living thing. This kind of nested categorization underpins much of how people organize knowledge and make inferences.
Laboratory research has shown that hierarchical classification can be established as a form of relational framing. In one study, researchers used training procedures to establish arbitrary symbols as cues for “member of” and “includes” relations, then tested whether participants could derive new hierarchical relations they had never been directly taught. The results showed patterns consistent with what cognitive developmental researchers describe as transitive class containment and property induction, meaning participants treated the hierarchies as genuinely nested and transferred properties from the broader category to the member.
This work has moved beyond the lab. A study with children with autism used hierarchical relational training and then tested whether participants could apply what they learned in a game of “I Spy.” After training on categories and their members, all participants successfully identified stimuli when given verbal clues that combined two category names, demonstrating a transformation of function, the ability to use learned relations in a new context.
Even more foundational work has examined how to establish the building blocks of categorical and distinctive (difference-based) relational responding in children with autism who had not yet developed these skills. Training was effective in establishing the target responses, and the skills generalized to untrained stimuli, which is exactly the kind of transfer that makes RFT-based intervention appealing in the first place.
How Widely Adopted Is RFT Among Practicing Behavior Analysts?
Despite the growing evidence base, RFT remains underutilized in everyday ABA practice. Surveys of Board Certified Behavior Analysts (BCBAs) tell a consistent story: practitioners are interested but underprepared. In one survey, about 63% of BCBAs acknowledged that RFT falls within the scope of ABA, and over 90% expressed interest in learning about it. But the actual uptake is low, with respondents citing insufficient training as a major barrier.
A second survey painted a similar picture but with sharper numbers on the education gap. While about 78% of behavior analysts reported receiving formal education in equivalence-based instruction (a related but older framework for derived relations), only about 16% reported formal education in RFT specifically. The authors suggested that adoption of RFT-derived procedures may be directly proportional to formal education, and that a lack of accuracy on basic RFT knowledge questions pointed to a real capacity gap in the field.
This is worth lingering on. The gap is not one of hostility or rejection. Most practitioners see RFT as relevant and want to know more. The bottleneck is practical: graduate programs in behavior analysis have been slow to incorporate RFT into their curricula, and the clinical translation of RFT concepts into step-by-step intervention protocols has lagged behind the basic science. For parents choosing an ABA provider, this means that “RFT-informed” programming is available but not yet standard. Asking a provider whether they use RFT-based curricula like PEAK, or whether their programming targets derived relational responding, is a reasonable way to gauge whether a program goes beyond traditional discrete trial approaches.
The Connection to Acceptance and Commitment Therapy
RFT did not originate within ABA. It grew out of the same research program that produced Acceptance and Commitment Therapy (ACT), a psychotherapy approach focused on psychological flexibility. ACT is, in a sense, RFT applied to adults dealing with anxiety, depression, chronic pain, and other challenges. Both ACT and RFT treat human suffering as partly a product of how language and relational responding can trap people in rigid, unhelpful patterns of thought.
This connection has created an interesting professional tension. ACT is widely used in clinical psychology but has only recently been discussed as something that might fall within an ABA practitioner’s scope of practice. One paper argued that ACT is fundamentally a behavior-analytic approach to changing verbal behavior for the purpose of producing meaningful overt behavior change, and offered guidelines for how behavior analysts might incorporate it into their work. Survey data show that a slim majority of BCBAs (about 53%) agree that ACT is within ABA’s scope, though half do not consider it part of their own personal practice.
For ABA practitioners who work with older children, adolescents, or adults, the ACT-RFT connection opens up possibilities for addressing issues like rigid rule-following, avoidance of uncomfortable feelings, and difficulty adapting to changing circumstances. These are challenges that traditional ABA programming, focused as it is on skill acquisition, does not always address directly.
Rule-Following and When It Becomes a Problem
One area where RFT has offered a genuinely new lens for ABA is in understanding rule-governed behavior. People follow rules all the time: “Look both ways before crossing the street,” “Study hard and you’ll get good grades,” “If you feel anxious, avoid the situation.” Rules are efficient because they let you skip trial-and-error learning. But rules can also become problems when they are inaccurate, outdated, or applied too rigidly.
RFT researchers have argued that rule-following is itself a form of derived relational responding, and that understanding it this way opens up more precise ways to intervene when rules become maladaptive. Research connecting rule-governed behavior to derived relational responding has explored whether it might be useful to think of rules not just as verbal instructions but as networks of derived relations whose influence depends on the relational context they are embedded in.
This has clinical relevance. A study with adolescents who had depressive symptoms found that while accurate instructions led to quick and efficient learning regardless of depression levels, inaccurate instructions led to problematic rule-following specifically in the group with high depressive symptoms. These adolescents got stuck following rules that no longer matched reality. The effect got worse when participants were also given what the researchers called superordinate pliance instructions, essentially, rules about following rules (“always do what you’re told”).
For ABA practitioners, this research highlights a real risk: if a child or adolescent has been extensively trained to follow instructions (which is a core part of many ABA programs), they may become overly rule-governed and struggle to adjust when circumstances change. RFT provides both the conceptual framework for understanding this problem and potential strategies for addressing it, such as training flexible relational responding and contact with direct consequences rather than verbal rules alone.
What Brain Imaging Tells Us About Derived Relations
A strand of RFT research has used neuroimaging and electrophysiology to investigate what happens in the brain during derived relational responding. Brain imaging studies have found that when people make inferential judgments based on derived relations (for instance, figuring out that if A equals B and B equals C, then A must equal C), the anterior hippocampus on both sides of the brain shows activation. Simpler symmetrical relations (if A equals B, then B equals A) engage the parahippocampus instead, while processing foils, or incorrect pairings, activates broader frontal and parietal regions.
Separately, research using event-related potentials (ERPs, which measure electrical activity in the brain with millisecond precision) found that different types of analogical reasoning produce measurably different brain signatures. When people responded to “similar-similar” analogies (A is to B as C is to D, where all relations are sameness relations), they were faster and showed distinct left-hemispheric prefrontal activity compared to “different-different” analogies. The researchers interpreted this as evidence that relationally simpler patterns are processed differently at a neural level, which supports the RFT prediction that relational complexity is a real, measurable dimension of cognition.
This kind of neuroscience work is still in its early stages and does not directly change what happens in a therapy room. But it does something that matters for the field’s credibility: it provides biological evidence that the relational processes RFT describes are not just theoretical constructs. They correspond to observable, distinct patterns of brain activity, which strengthens the case for taking RFT seriously as a scientific framework rather than treating it as an abstract philosophical exercise within behavior analysis.